April 16th 2026 | 2min read

Using Linguistics to Uncover AI Chatbot Addiction

Following our recent exploration of emerging safeguarding trends, this spotlight takes a closer look at one area of focus: AI chatbot addiction.

As technology continues to shape how young people communicate and seek support, understanding how these interactions are experienced and described provides valuable insight into behaviour and potential safeguarding concerns. In this spotlight, Dr Charlotte-Rose Kennedy, Senso’s Safeguarding Language Specialist, shares insights from her recent research exploring how individuals describe their experiences with AI chatbots, and what these patterns may reveal.

AI Chatbot Addiction: A Rapidly Growing Safeguarding Concern

AI chatbots have rapidly become part of students’ everyday digital lives, from homework assistance to entertainment and social interaction. These platforms are designed to help users fulfil social needs and facilitate emotional bonds by emulating human-like attributes. As a result, users can begin to perceive them as real people and form friendly or even romantic relationships with the bots.

For safeguarding teams, this creates a new challenge. AI chatbots are not necessarily designed with children's safety in mind. Research from Internet Matters found that AI chatbots can produce inaccurate or unsafe responses and expose children to harmful or age-inappropriate content. Meanwhile the Internet Watch Foundation (IWF) has reported cases where AI chatbots have been used to simulate ‘abhorrent’ sexual scenarios with children.

In this spotlight, Dr Charlotte-Rose Kennedy, Senso’s Safeguarding Language Specialist, shares insights from her recent research exploring how people describe their experiences with AI chatbots - and what these patterns may reveal about the potential risks associated with over-reliance on AI companions.

Analysing Language Patterns in AI Chatbot Use

Dr Charlotte-Rose Kennedy analysed a 284,833-word dataset of forum posts written by individuals discussing attempts to reduce or quit AI chatbot use. By examining how people used the phrase “I feel”, the research revealed important insights into the motivations, experiences and emotional impact associated with AI chatbot use.

Three clear themes emerged from the dataset:

• Why people turn to AI chatbots
• The negative consequences of chatbot addiction
• The challenges people face when trying to stop using them

Why People Turn to AI Chatbots

One of the clearest themes in the dataset was that many people turn to AI chatbots to fulfil social and emotional needs that feel unmet in real life.

Users frequently described feelings of loneliness, isolation, or a desire for companionship. Others described using chatbots as a way to cope with difficult emotions or escape from real-world pressures.

While exact extracts from the dataset are not included in this blog to protect the anonymity of the forum posters, examples from the dataset included phrases similar to:

“I have nobody to talk to and I feel so alone”

“I feel safe and loved when I'm with them”

“When I feel depressed, I use it to escape reality”

These findings provide valuable insight into why AI companions can become appealing - particularly for young people who may already be experiencing loneliness or emotional distress.

The Negative Impact of AI Chatbot Addiction

The research also highlighted many users expressed feelings of shame, declining mental wellbeing, and a sense that they were unable to stop using the technology. Some users also described experiencing declining cognitive engagement, including difficulties concentrating, writing, or maintaining meaningful real-world relationships.

Challenges When Trying to Stop

Another theme that emerged was the difficulty people experience when trying to stop using AI chatbots. Because many users had formed emotional attachments to the technology, attempts to quit often produced feelings similar to losing real-life friends.

Examples from the data alluded to missing friends, feeling lonely, and having urges to return to the app. Some users described symptoms such as mental exhaustion, low motivation, and depression during attempts to stop using the platforms.

Why This Research Matters for Safeguarding

Research like this helps us identify why people engage with AI chatbots and how those relationships can develop over time.

By identifying the language patterns associated with these behaviours, we can continue refining the safeguarding libraries used within Senso’s monitoring and filtering tools. This allows schools to be alerted when language patterns start to suggest a student may be experiencing distress, isolation, or developing unhealthy relationships with technology, supporting earlier intervention and more informed safeguarding decisions To find out more, read our full article here.